I l R d T h lImpulse Radar Technology –Fundamentals and Applications

DAG T. WISLAND, CEO +47 913 67 679DAG T. WISLAND, CEO +47 913 67 679dag@novelda.nodag@novelda.no WWW.NOVELDA.NOWWW.NOVELDA.NO

OUTLINEOUTLINE

Novelda Company BriefNovelda Company BriefUWB Radio FundamentalsN ld I l R d T h lNovelda Impulse Radar TechnologyRadar Application ExamplesRecent Developments CMOS RadarConclusionConclusion

NOVELDA COMPANY BRIEFNOVELDA COMPANY BRIEF

NOVELDA ASNOVELDA AS

A fabless semiconductor company specializing in Nanoscale p y p gwireless low-power technology for ultrahigh-resolution impulse radar

Developing CMOS impulse radio standard components, as well as Application Specific Integrated Circuitspp p g

Applications for our technology spans a wide range of areas from medical and industrial high precision sensors to personalized wireless healthcare and more

COMPANY BACKGROUNDCOMPANY BACKGROUND

Founded in September 2004 by:Dag T. Wisland – CEO/Associate professor Univ. of OsloTor Sverre (Bassen) Lande – Professor Univ. of OsloEinar Nygård – Industrial IC management / EntrepreneurEinar Nygård Industrial IC management / EntrepreneurEirik Næss-Ulseth – Business developer / Entrepreneur

C b iCore business:Focus on impulse radio technology in CMOSProduct/Market: Low-energy, short-range, high-precision impulse radarR&D driven development

EU/RCN/IN projects

NOVELDA EXPERTISENOVELDA EXPERTISE

Micro-/nano electronics and electrical engineering 17 employees - 3 Ph.D, 11 M.Sc., 3 B.Sc.

Advisory board - 2 Professors

Management group with up to 28 years R&D experience

International Cooperation with several acknowledged p guniversities and research facilities

UWB RADIO FUNDAMENTALSUWB RADIO FUNDAMENTALS

HISTORY OF RADIOHISTORY OF RADIO

Everything started with sparks!Heinrich Rudolf Hertz

Wireless with sparks (1886-89)

Guglielmo MarconiGuglielmo MarconiLong distance radio

> 1.5 km in 1895

Nikola TeslaFather of radio telegraph (patent 1891)

It all started with the sparkIt all started with the spark…

WHERE DID THE SPARK GO?WHERE DID THE SPARK GO?

Dominant in these early daysTransatlantic telegraph (Morse)Hard to share

Si ifi t i t fSignificant interference

Carrier based radio (1910)Coding on top of narrowband carrierCoding on top of narrowband carrierSharing by coordinating carrier frequencies

Still the way radio worksy

Time Frequency

IMPULSE ↔ CARRIERIMPULSE ↔ CARRIER

Carrier based radio

Time FrequencyNarrow frequency bandCarrier → no information, but significant energy

Impulse radio

q y

Impulse radio

UWB

Wide frequency band

FrequencyTime

UWB

No power demanding carrier

WHY IMPULSE RADIO NOW?WHY IMPULSE RADIO NOW?

Legal transmission!FCC in 2000/2002

3.1-10.6GHz7 5Ghz wide band7.5Ghz wide band

Low energy emissionEIRP<-41.3dBm/MHzClose to noise…

Significant signal interference

Wireless networksWiFi 802.11a

Largest unlicensed frequency band ever released!

EUROPEAN / ASIAN RESTRICTIONSEUROPEAN / ASIAN RESTRICTIONS

6.0 GHz – 8.5 GHz

12

IMPULSE RADIO FEATURESIMPULSE RADIO FEATURES

Different from narrow band radioTime domain processing

No mixers like standard radio (frequency)No carrier More channels easy (traded for bandwidth)Different penetration propertiesDifferent trade-offsDifferent trade offsTechnology friendly

Speed – virtue of modern technology

New featuresHigh accuracy rangingShort range radar

IMPULSE RADARIMPULSE RADAR

Impulse radarOld technologyHülsmeyer patent

World War II developmentUSA, Soviet, Germany, EnglandF b dFrequency based systems

Ground Penetrating Radar (GPR)I d t d d fIndustry and defense

Look into groundHard to make

NOVELDA RADAR TECHNOLOGYNOVELDA RADAR TECHNOLOGY

FROM RESEARCH TO PRODUCTFROM RESEARCH TO PRODUCT

IdeaMarket need

Samuel Morse – Pulse coding

e eedEnabling technologyPeople – Competence

R&DR&D

Heinrich Hertz – Pulse comm. Novelda NVA6100Nanoscale Impulse Radar

Soft fundingVC funding

Subcontractors Nanoscale Impulse Radar=

Competitive advantages

Guglielmo Marconi – Radio system

MICROPOWER IMPULSE RADARMICROPOWER IMPULSE RADAR

Tom McEwan (1994)Lawrence Livermore National Laboratory (LLNL) y ( )

Integrating Peak DetectorAnalog lossy integrator

i l lSingle samplerNoise-like pulsesUnavailable inlow voltage digital CMOS

E. M. Staderini Medical RadarHome made McEwan radarHome made McEwan radar

First Novelda proof-of-concept (2005)

17

ENABLING TECHNOLOGYENABLING TECHNOLOGYQ iQuestion:

How to capture electromagnetic pulses traveling with speed of light achieving millimeter spatial resolutionKeep a small physical size and low unit costKeep a small physical size and low unit cost

Answer: Nanometer CMOS technology (<90 nm)

Silicon barMaterial technology

Final IC productAdvanced packaging

Nanoelectronic system design (IP)Advanced CMOS production

IMPULSE RADAR PRINCIPLE (2)IMPULSE RADAR PRINCIPLE (2)

Tx

Rx

T itt dTransmittedPulse (Tx)

ReceivedPulse (Rx)

PENETRATION ABILITIESPENETRATION ABILITIESPulses penetrate heavy matter

Body penetrationLayered structure

Reflection and penetration due to resonanceAlways some frequency with wavelengthAlways some frequency with wavelength proportional to layer thickness

Impulses are wide band covering large frequency range

20

CTBV DOMAINCTBV DOMAINTime domain processing

Exploring technology speed90nm → 12ps inverter delay

Binary for low supply voltage1V → binary values

Continuous Time – Binary Value

Discrete Continuous

Time

DigitalCTBV

Neuromorphic/spike

Analog

Discrete Continuous

(switched cap)Analog sampled−data

Binary

Continuous

Value

No high speed clockPower efficient

i i lik i fi i l k

g(switched−cap)e

Continuous time like “infinite” clock

21

CTBV IMPULSE RADARCTBV IMPULSE RADAR

Single chip CMOS impulse radar – 1mm2 siliconHigh speed sampler

Millimeters resolutionSingle pulse → multiple depth

128 parallel digital integratorsp g g256 in 2. generation

Covering ≈ 27 cm depthMultiple ranges (unique feature)

Power efficient1.1V supplyNo high speed clock

DelayRange

strobeg p

Low speed SPI readout≈ < 30mA total

Inputstage

>60 GHz sampler

256 x24 bit samplers

Serialout

22

256 x24 bit samplersSPI

CTBV PULSE GENERATORCTBV PULSE GENERATOR

Dual slope generatorP ffi i

Measured dual slope

Power efficientNo stand-by power

N ill t / l k V)No oscillator/clock

Limited signal swing Out

put (

V

Time (s)

90nm STMicroelectronics process50 Ω load with cable to scope

Period ≈ 300 ps

Time (s)

Period 300 ps

HIGHER ORDER GAUSSIANHIGHER ORDER GAUSSIAN

Longer cascadeScaling sizes

• No oscillator • Works in digital CMOS• No stand-by power

Works in digital CMOS• Limited signal swing

RADAR PERFORMANCERADAR PERFORMANCE

Reflected energy is low!Buried in noiseOnly recoverable with heavy integration

1e+08Desired rms error = 0.001, 0.00316, 0.01 (uppe r, middle, lower graphs)

100000

1e+06

1e+07

red

(n)

100

1000

10000

ampl

ings

requ

ir

In noisy environments Swept threshold sampling is

approaching an ideal integrator in performance!

0.1

1

10

S

Stochastic resonanceSwept thresholdAnalog average

25

0.1 1e-04 0.001 0.01 0.1 1 10

Input noise(σN)

NVA6XXX NANOSCALE IMPULSE NVA6XXX NANOSCALE IMPULSE RADARRADARRADARRADAR

Single chip Impulse RADARClose Range Operation 0-60mHigh-resolution, millimeter range

Sub mm with interleaved samplingSub mm with interleaved sampling

High speed > 30GHz sampling rateDepth perception, 512 simultaneous depthsLow energySmall size, CMOSTX frequenciesq

0.7 GHz – 2.4 GHz (Medical)3.1 GHz – 5.6 GHz (US market)6.0 GHz – 8.5 GHz (EU/Asian market)( )

MEASURED TX FREQUENCY MEASURED TX FREQUENCY TUNABILITYTUNABILITYTUNABILITYTUNABILITY

MEASURED TX/RX SPECTRUMMEASURED TX/RX SPECTRUM

MEASURED TIME DELAY MEASURED TIME DELAY TEMPERATURE SENSITIVITYTEMPERATURE SENSITIVITYTEMPERATURE SENSITIVITYTEMPERATURE SENSITIVITY

MEASURED TX TEMPERATURE MEASURED TX TEMPERATURE SENSITIVITY (MED VERSION)SENSITIVITY (MED VERSION)SENSITIVITY (MED. VERSION)SENSITIVITY (MED. VERSION)

RADAR APPLICATION EXAMPLESRADAR APPLICATION EXAMPLES

NVA R6XX DEVELOPMENT KITS NVA R6XX DEVELOPMENT KITS

NVA R6XX DEVELOPMENT KITS NVA R6XX DEVELOPMENT KITS

Common featuresCl R O i 0 60Close Range Operation, 0-60mHigh-resolution Simultaneous observation of 512 depths with programmable depth resolutionHigh speed; > 30GHz Sampling rateUltra low RF emission (< FCC Part 15 limit)GUIC-library with API, Matlab examples

Two hardware modulesReference RF design with Digital SPI interface and SMA connectors for RX and TXIO module featuring a Micro controller with USB 2.0 Full speed and JTAG interface

NVA R6XX DEVELOPMENT KIT NVA R6XX DEVELOPMENT KIT VERSIONSVERSIONSVERSIONS VERSIONS Novelda R620 – Q4 2010:

Single chip CMOS NVA6000P Impulse RADARSingle chip CMOS NVA6000P Impulse RADARTransmit bandwidth (-10dB) from 6 GHz to 8.5 GHz Designed for ETSI/FCC compliant

Novelda R630 – TBA:Single chip CMOS NVA6100P Impulse RADARTransmit bandwidth (-10dB) from 0.7 GHz to 2.4 GHz

Novelda R640 - Released:Single chip CMOS NVA6100P Impulse RADARSingle chip CMOS NVA6100P Impulse RADARTransmit bandwidth (-10dB) from 3.1 GHz to 5.6 GHz Designed for FCC compliant

Novelda R650 – TBA:Single chip CMOS NVA6000P Impulse RADARSquare pulse with adjustable pulse-width from <100ps to >1ns

APPLICATION EXAMPLESAPPLICATION EXAMPLES

Monitoring vital Signs; breath, pulse,

Gas/fluid Level detection/gauging

Energy automationSnow measurements

Soldier monitoringSurveillance

blood pressure, stress, sleep, etcDiagnostic sensors

AutomotiveInspection

Structures Line of sight andthrough the wall

APPLICATION EXAMPLEAPPLICATION EXAMPLEREMOTE PULSE MEASUREMENTSREMOTE PULSE MEASUREMENTSREMOTE PULSE MEASUREMENTSREMOTE PULSE MEASUREMENTS

Non-invasiveMeasures mechanical movement

Improved diagnostics qualityEarly sign disease detectiony g

Skin/Fat/Tissue penetration Ultra Wideband Radio

Low cost (Single IC)Low cost (Single IC)PhysiciansSport/Leisure

Low energy / Small sizeLow energy / Small sizeLong battery lifetimeSport watch

RECENT DEVELOPMENT CMOS RECENT DEVELOPMENT CMOS RADARRADAR

RECENT CMOS RADAR DEVELOPMENTRECENT CMOS RADAR DEVELOPMENT

JSSC 2010JSSC 201077 GHz FMCW (Toshiba Corp.)

ISSCC 2010ISSCC 201077 GHz FMCW (Nat. Taiwan Univ.)

IEEE Radar Conference 201077 GHz FMCW radar (Univ. of Melbourne)

VLSI Symp. 200977 GHz FMCW radar (Toshiba Corp.)( p )

CONCLUSIONSCONCLUSIONS

“Real UWB” impulse radar is feasible inReal UWB impulse radar is feasible in standard CMOS using the CTBV approachmm precision ranging is achievablemm-precision ranging is achievableImpulse radar penetrates human tissueCalibration mechanisms must be employed to cope with delay variation Commercial UWB is not dead!

ACKNOWLEDGEMENTSACKNOWLEDGEMENTS

Thanks to RCN for funding our R&D throughThanks to RCN for funding our R&D through the “UWBPOS” and “CARDIAC” projectsThanks to “Bassen” for providing nice slidesThanks to Bassen for providing nice slides on UWB/Impulse radar fundamentalsTh k N l i G D fThanks to Nanoelectronics Group at Dept. of Informatics, Univ. of Oslo for fruitful R&D

icooperation

THANK YOU FOR YOUR ATTENTION!THANK YOU FOR YOUR ATTENTION!

QUESTIONS?QUESTIONS?